On Thu, 10 Sep 2026 17:35:48 GMT, Kangcheng Xu <[email protected]> wrote:
>> Hi, >> >> This PR adds intrinsics to `Preconditions.checkFromToIndex()` and >> `Preconditions.checkFromIndexSize()` to produced optimized IR that uses >> `RangeCheck` node instead of implicit comparisons. >> `Preconditions.checkIndex()` is also refactored (without additional >> optimizations) to use the same helper function. >> >> Some common patterns where calling `checkFromToIndex` or >> `checkFromIndexSize` in a loop can have range checks in main loop eliminated >> completely and, therefore, enables empty loop removal. >> >> IR and correctness tests are included and passing. Additional tests on >> `Preconditions.checkIndex()` were also added. >> >> Thanks! >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Kangcheng Xu has updated the pull request incrementally with two additional > commits since the last revision: > > - simplify logic by allowing deopt with size == 0 > - remove left over code The latest commit simplifies bound check logic by check the first and last indices of a subrange. This avoids the `u<= length` to `u< length + 1` workaround. The old code had a corner case causing deopt with `size == jint_max` or `size == jlong_max` . The new corner cases are deopt-ing with `size == 0` or `to - from == 0`, which are equally unlikely to happen in practice. the latest benchmark is the following: <img width="2681" height="889" alt="image" src="https://github.com/user-attachments/assets/d05ab241-f702-4939-8b75-5192373509fc"; /> It is slightly more noisy on array copies, but no overall improvement or regression compared to before the simplification. ------------- PR Comment: https://git.openjdk.org/jdk/pull/31138#issuecomment-5624110331
